Those fearing imminent mass unemployment driven by robots and AI should be following stories like this. They also should be looking at the actual productivity numbers.
See also the details of work and the realities of automation.
21 thoughts on “Robot Gets Hired, Tries Hard, But Can’t Do the Job”
Comments are closed.
It is tough being at the cutting edge. It sounds like Boeing had earlier had a more modest level of automation which enhanced productivity. Then the evil Germans (in league with Airbus???) sold them an expensive pup on full automation.
From a position of complete ignorance, it would seem there is a world of difference between the appropriate kind of automation in a car factory which turns out 250,000 vehicles per year and in a wide-body aircraft plant turning out about 100 planes per year. Low volume and flexibility may require more of a human-assistance type of automation than a completely machine-driven form.
Boeing is at the bleeding edge, and set-backs are to be expected. But some would argue that Boeing’s management has spent too much time & effort over the last few years thinking about stock buy-backs instead of thinking about technology.
Boeing appears to have serious management problems. This is a national security issue given the concentration in the defence industry. I suspect there is much rot in all of it that will not be revealed until one year into the next real war.
Boeing has been riveting their airplane together forever, you’d expect that they would be able to judge what would work and what wouldn’t. Eventually, what happened will leak out.
Here’s a link to show what they’re talking about:
https://www.youtube.com/watch?v=GevBp3nj878
The one in the article wouldn’t work for me for some reason.
Trying to go to 100% automation today is foolish, and so obviously going to fail that one wonders if there was some sort of agenda going on. Even Amazon’s warehouses still don’t have a replacement for the human hand, as far as I’ve read, though everything else is automated.
Here are a few observations from working for a trucking company at a dedicated customer account. The warehouse has seen worker productivity drop about 25% over the past several years despite implementing various technologies. According to a supervisor, this is a result of high turnover. Previously, the average tenure was several years but is down to almost nothing. This is a well-known European industrial company that starts warehouse workers at about $40k with European type benefits right off the bat. The new workers just don’t stick around. At 10 years, they would probably be over $60k with overtime.
Employee retention is what is going to drive automation in certain sectors more than anything. There are certain warehouse positions that are difficult to staff due to the processes involved being highly repetitive and mundane. At another warehouse of an online retailer, people would essentially refuse placement by early outing. Eventually, they would be terminated for point accumulation or just quit. This place had a 90% turnover rate. If they could automate certain positions, it could drastically cut down on turnover. In articles about a new warehouse being automated, this is often only briefly mentioned as the driving force. Not too long ago, an article discussed a new warehouse being automated but it would still employ 400 people.
As far as trucking, the only place possible to see any sort of self-driving OTR in the next 20 years is in the I-10 corridor. The older post from Marginal Revolution is accurate. Much of the work centers around compliance in one form or another.
My guess is that most people reading this blog would be shocked at how poorly technology is implemented in many manufacturing and logistics facilities. Places like Boeing and Amazon get nice articles written about them but they are the exception.
“My guess is that most people reading this blog would be shocked at how poorly technology is implemented in many manufacturing and logistics facilities”
Nah. I’m sure it wouldn’t surprise us in the least. I’d offer that you’d be shocked at how terribly tech is implemented by the DoD when it comes to imagery, comms, etc, but probably not. It’s a miracle anything functions at all.
Bob: “According to a supervisor, this is a result of high turnover. Previously, the average tenure was several years but is down to almost nothing.”
Management guru Peter Drucker warned — “Beware the recurrent crisis!”. If turnover has been accelerating and new hires are quitting rapidly, the problem is not in the workforce, it is in the management. It would be a nice problem to fix.
Start by shaking down the Human Resources group, probably led by a supercilious Euro import and staffed by a bunch of college-credentialed females who spend their time on their phones when they are not bitching about their current boyfriends. Do they have coherent proposals for how to improve the situation? Do they even bother to do exit interviews with the workers who quit?
Then take a serious look at policies applied in the warehouse — little things like how bathroom and smoking breaks are handled can make a big difference in whether this is a place human beings want to work. What kind of training & support do new hires get? Associated with this, a very serious look at the supervisors. Do increases in turnover happen to coincide with changes in supervisors? Do the supervisors have ideas about how to improve retention? If so, who in management is preventing them from implementing those ideas?
After all the management, supervisory, and policy problems have been fixed and the workplace has been made more attractive, look at the pay scales.
“Employee retention is what is going to drive automation in certain sectors more than anything.”
Sometimes if seems that poor managers look to automation as a way to avoid having to face up to their own incompetence in dealing with human beings. Boeing may be a case in point — facing poor industrial relations in Seattle, executive management ran away to Chicago.
In her book ‘The Good Jobs Strategy’, author Zeynep Ton cites some really appalling numbers about inventory inaccuracies in the systems of major chain retailers…which systems, of course, drive ordering and hence have a big influence on stock-outs and improperly balances inventory investments. Plenty of money has been spent on these technology investments, but her analysis is that bad work-process design, combined with understaffing and poor employee motivation, greatly reduces their effectiveness. An interesting book, which I plan to review one of these days.
Also, see this story about Target Canada:
https://chicagoboyz.net/archives/51968.html
But some would argue that Boeing’s management has spent too much time & effort over the last few years thinking about stock buy-backs instead of thinking about technology.
This is the sort of thing that almost killed Ford 30 years ago.
It probably did kill GM. The story of the Saturn plant is in the book, “Crash Course.”
I remember a story a few years ago about I-Phone production in China. Part of it was about what I’ll call the Cute Chinese Girl, CCG, at the end of the line that placed the phones in their final packaging.
That was not all she did. First she turned it on as a final test of operation, then she used a wipe of some sort to remove any smudges of fingerprints. She examined it closely for any sort of damage like a scratch. Finally, she turned it off, placed it in a poly bag and into the box with all the accessories and booklets. She did this in what almost seemed a single continuous motion for, I assume, 8-10 hours a day.
I considered what it would take to automate all of this and concluded that it was probably possible. The challenge would be to design a machine that could be reconfigured from I-Phones to Mac-Books as easily as the CCG.
At the time, the CEO of Foxcon made some grandiose promises/threats about adopting automation. None of which have come to pass. They instead seem to be moving on in search of lower cost bulk labor.
When it comes to automating transportation, there is already a system that allows a three person crew to haul hundreds of truck loads at once burning a fraction of the fuel. When you eliminate all the reasons that long haul trucking still thrives in the face of competition from efficient railroads, you’ll have something you can automate and probably also have eliminated the justification for automating individual trucks at the same time.
“But some would argue that Boeing’s management has spent too much time & effort over the last few years thinking about stock buy-backs instead of thinking about technology.”
Quite possibly, but I would certainly hope that decisions on work processes and robotics for airliner manufacturing are not requiring extensive CEO & Board involvement. Unless the amounts involved are in the billions rather than in the tends of millions…I think I saw $20MM or so cited somewhere…this should be a matter for the head of Boeing Commercial Airplanes, at the most.
“FAUB is a horrible failure,” one mechanic told The Seattle Times in 2016. Another called the system “a nightmare” that was snarling 777 production. Yet Boeing insisted then that these were teething pains that would pass.
Might be some principal–agent problems.
David Foster: You have probably hit the nail on the head. The ability to efficiently build air frames is the whole company. That this sort of “bet-the-company” situation didn’t reach the top until it was a disaster is apparent in the 737-MAX debacle as well. Do you want to fly in a plane built by a hedge fund?
MCS….actually, I was making a different, and somewhat contrary, point. The decision to implement a new assembly technology for one model of airliners within the commercial airplane division *should not* be a matter for the Boeing CEO and board.
Boeing Commercial Airplanes is a $60B business, larger than many F500 companies. It is run by someone whose title is President and CEO (of the business unit.) As near as I can tell, he has manufacturing as well as engineering reporting to him. (There is also a SVP of manufacturing reporting at the top level, but I think she is probably running more of a staff function cutting across the businesses)
There’s always a temptation, when bad things happen, to centralize….but eventually, decisions can get tied up in endless knots and other kinds of bad things happen. And is there really any reason to believe that a better decision on the assembly robot would have been reached at a higher organizational level? Indeed, the decision to build and try this robot may eventually turn out to be the *right* one from a knowledge-gaining point of view.
The 737 Max issue is somewhat different, because of the extreme penalty clause ($1MM/airplane) that was included in the Southwest contract if additional simulator training should be required, and the rather bizarre decision not to document that system in the pilot manuals.
I’m pretty sure I saw your point, I just didn’t agree with it. I happen to think that Boeing’s ability to make airplanes that don’t crash should rise to the attention of the CEO. What, exactly, could he be doing that’s more important. That million per plane looks pretty cheap now.
“I happen to think that Boeing’s ability to make airplanes that don’t crash should rise to the attention of the CEO.”
But practically everything Boeing *does* affects their ability to make airplane that don’t crash…an exaggeration, but not all that much of one. Decision on flaps-down stall speed?…affects landing speed and runway distance required, hence impacting how much runway margin will be left before running off the end of the runway when things go wrong. Computer language used for implementing critical avionics and flight control code?…quite likely can impact the possibility of errors creeping in. Criteria for hiring test pilots?…Hundreds of thousands of decision impacting flight safety, many of them impacting cost and delivery as well. What should be the decision run for which ones need to be personally reviewed by the CEO?
And lots of similar decisions on the Boeing Defense Products side, along with many other impacting combat effectiveness…itself a life-and-death issue.
Anyhow, I don’t think the assembly robot decision was really about safety, it was more a pure efficiency and cost issue, possibly with some impact on delivery schedules.
MCS: “I happen to think that Boeing’s ability to make airplanes that don’t crash should rise to the attention of the CEO.”
It certainly is hard to disagree with that. To David F’s point, of course the CEO cannot (and should not) personally check every line of code and inspect every rivet — but it is the CEO’s job to set up the organizational structure and choose the right staff to make sure that someone does. If there are blockages in the information flow between the line workers and the decision makers (which certainly seems to have happened in the case of the 737 MAX), that is most certainly the CEO’s responsibility.
So what were Boeing’s executives doing that took their focus away from the extremely complex process of manufacturing heavier-than-air flying machines? If you look at Boeing’s Annual Report, it appears that “diversity” was getting more attention than mere competence. And then there is this:
https://www.epsilontheory.com/when-was-i-radicalized-boeing-edition/
“In 2017 alone, the one good stock-performance year Boeing has had in a decade, $3.8 billion went to Boeing management and directors. That’s 29% of cashflow from operations for the year.”.
As CapitalistRoader says — principle-agent problems.
It’s not an exaggeration to say that every employee affects the ability of Boeing to build safe and economical airplanes, it’s the literal truth. Of course, no one person can know and see all that is necessary to manage something this size. What has to happen is that a system is in place that competent people can use to insure that the unavoidable mistakes don’t grow out of control. That’s why they pay him the big bucks.
The assembly issue, one hopes, never becomes a safety issue because someone was paying close enough attention to spot the problems. Six years is a very long time to solve a problem and you have to wonder if all of the problems on all of the planes were caught and fixed or if some could be waiting for enough load cycles to manifest.
Any of these and other issues in isolation would be troubling but there seems to be a pattern that implicates the entire organization. Proper judgements aren’t being made, which indicates that the right people aren’t in the right positions and that the wrong priorities are being set, with the short term overbalancing what should be the bedrock foundation that the company was built on.
By the way, the point about the Southwest penalty that I hadn’t heard about before doesn’t reflect well on Southwest. I’d love to hear them try to reconcile that with their “Safety Culture”.
The $1MM penalty per airplane seems disproportionate to the costs of sending pilots to do sim training. I bet that SW would have still done the deal if the amount had been, say, $100K per airplane instead of $1MM, and the pressure to make everything look/seem the same would have been a lot less.
““In 2017 alone, the one good stock-performance year Boeing has had in a decade, $3.8 billion went to Boeing management and directors”
To actually measure this number accurately, one would need to look at the proxy statement and see what was actually awarded and its value. Looks like they’re all online, if anyone would like to dig into them: here’s the one for 2018:
http://s2.q4cdn.com/661678649/files/doc_financials/annual/2018/2018-Proxy.pdf